Skip to main content
Welcome guest. | Register | Login | Post

Building an OEM Solution

9 replies [Last post]
supermike's picture
Offline
Joined: 2006-02-17

I have a long commute to and from my dull day job. I'm trying to think out some ideas. I'm working on this killer PHP app on Linux that I may release next year. Eventually as one of the ideas I'm throwing around, I'd like to make a special bundle so that people could purchase the web app and a Linux OS, pre-installed, on a 2U rackmount system.

Has anyone dabbled in purchasing a generic OEM rackmount system and put your own custom bezel button on it, put Linux on it, and then either custom-built the web admin interface for it or found one and customized it?

* What's the cheapest, most-sensible RAID-5 config out there?

* I'm interested in what you purchased, how much it cost, and whether you felt it was worth it or could make other recommendations about it.

* I'm wondering whether or not you found a great web admin interface that you could customize, or had to make one from scratch. When I talk web admin, I mean to do things through the web browser (for noobs) like change the IP address, bounce the server, start/stop daemons, etc.

libervisco's picture
Offline
Joined: 2006-05-04
Re: Building an OEM Solution
"supermike" wrote:

* I'm wondering whether or not you found a great web admin interface that you could customize, or had to make one from scratch. When I talk web admin, I mean to do things through the web browser (for noobs) like change the IP address, bounce the server, start/stop daemons, etc.

I haven't ever dealth with rackmounts and RAID configurations before, but I just thought I could respond with a humble recommendation for a web admin interface. You may have already known about it:
http://www.webmin.com/

It is modular and therefore extensible to whatever you need it to do. I imagine it can handle most of the administration tasks.

Offline
Joined: 2006-01-29

While I can't really give any firsthand experience with any OEm suppliers, I can give some info that may be helpful. I run a small systems admin/consulting/general Linux/OSS support and services company and had looked into possibly adding something similar to what you're thinking (minus the killer PHP app..lol).

The one thing I found is that many OEMs have rather large purchase minimums for such systems (as in per unit requirements), or require you to sign on for a minimum number of orders placed per quarter. Those who do not have such requirements seemed to run a bit higher in price, which cuts into profits. So for me, at this time it simply didn't make business sense to add a hardware component.

I would humbly submit that maybe you should think about adding a GUI to your PHP app (I know this may not make any sense since I have no clue about what your app is) for the admin interface. I would also suggest that maybe you would be better off concentrating on growing your client base with the app alone first before adding a hardware solution. Adding a hardware component to a primarily software business can be a huge adjustment once you factor in support costs, shipping to and from suppliers/customers, support time, failed part replacements, etc.

I wish you lcuk though. If you do go ahead with this, please post back and let us know what suppliers you went with, how it will be integrated to your existing business, etc. I wish you the best of luck with the app as well! Any hopes of discounts for Nuxified members :-D

supermike's picture
Offline
Joined: 2006-02-17
Re: Building an OEM Solution
"libervisco"][quote="supermike" wrote:

I just thought I could respond with a humble recommendation for a web admin interface. You may have already known about it:
http://www.webmin.com/

You can sort of tell that the main guy who started webmin was a big fan of Lotus Notes. It's got that tab/button kind of interface to it. This product sure looks handy, but dang if it doesn't need SKINS! Phew, what an eye sore.

Too bad it doesn't look as nice as eBox. (Although I don't care for green and orange, but that's entirely fixable.) I think eBox was started in Spain.

http://ebox-platform.com/index

supermike's picture
Offline
Joined: 2006-02-17
"dotmil" wrote:

<snip>... I run a small systems admin/consulting/general Linux/OSS support and services company and had looked into possibly adding something similar to what you're thinking...The one thing I found is that many OEMs have...<snip>

Dotmil, this was an entirely useful and thought-provoking post. I really do appreciate your insight. It looks like you've gone down this path here.

Now when you say build a GUI, do you mean build a PHP admin interface to administer the hardware, such as reboot it, set it to static or DHCP addressing, start/stop/restart services, etc.?

What is the app? It's sort of like Request Tracker (RT) on steroids and with a pretty interface, but still not entirely too complex. It's going to be marketed as both an as-is product and an SDK framework for someone to OEM off of me and customize per customer. I also plan to get my company and product (eventually) listed on GSA schedule with the USA Govt, so I have a feeling it will take off then.

I was thinking eventually when the product starts to sell, to do something to what SugarCRM did. They created SugarCube. They managed to get a nice rackmount server with a colored bezel, and then put a custom bezel button made and stuck on there. Inside, it's Debian, I think, along with SugarCRM pre-installed and ready to go. Of course, it's an expensive option to think about so early, but eventually I think it would sell well if my LLC (when I get that founded) were structured properly and the profits were coming in.

I was thinking along the lines of something as inexpensive as possible, initially, but with redundant power supplies, RAID-5 on 3 SATA drives, and 2 10/100/1000 NICs that can either be on separate IP addresses or aggregated on one IP to increase bandwidth. It would have Ubuntu Server on it and the /etc/apt/sources.list would only work from CDR that one purchased on subscription.

But this is entirely speculation at this point. Right now I'm in the stage of working the admin interface of the app itself. Everything else is written. Then, when that's done, I have to rebuild the guts of it so that it is an easier, consistent framework under the hood to deal with. And then there's the web-based install that I have to build for the app.

Offline
Joined: 2006-01-29
"supermike" wrote:

Dotmil, this was an entirely useful and thought-provoking post. I really do appreciate your insight. It looks like you've gone down this path here.

You're quite welcome for the comments; I learned much of this the hard way, trial and error; if I can help you skip some of the "error" portion of that, then its worthwhile :-)

Quote:

Now when you say build a GUI, do you mean build a PHP admin interface to administer the hardware, such as reboot it, set it to static or DHCP addressing, start/stop/restart services, etc.?

Yes, before I knew what the app was (and even now it could still be an option I suppose) I was thinking you could offer the additional functionality within th app itself, but include it only with the hardware edition or something.

Quote:

What is the app? It's sort of like Request Tracker (RT) on steroids and with a pretty interface, but still not entirely too complex. It's going to be marketed as both an as-is product and an SDK framework for someone to OEM off of me and customize per customer. I also plan to get my company and product (eventually) listed on GSA schedule with the USA Govt, so I have a feeling it will take off then.

Hey, another thing I have experience with (I'm a US gov't employee for my "real job")! Getting GSA listed will definitely broaden your potential client base. Still, consider you'll have to really "sell" your solutions as well. Uncle Sam's procurement policies aren't the simplest things in the world, so the vendors who do the best are those who make it easy to buy from them and will bend over backwards before and after the sale to support their products.

Quote:

I was thinking eventually when the product starts to sell, to do something to what SugarCRM did. They created SugarCube. They managed to get a nice rackmount server with a colored bezel, and then put a custom bezel button made and stuck on there. Inside, it's Debian, I think, along with SugarCRM pre-installed and ready to go. Of course, it's an expensive option to think about so early, but eventually I think it would sell well if my LLC (when I get that founded) were structured properly and the profits were coming in.

Yes, I think your idea is sound and makes sense, but I would definitely focus on build the app first and foremost. Get a few releases out, then talk to your client base to see if the hardware is osmething they would be interested in or not.

Best of luck, and keep me posted!

supermike's picture
Offline
Joined: 2006-02-17
"dotmil" wrote:

Getting GSA listed will definitely broaden your potential client base. Still, consider you'll have to really "sell" your solutions as well. Uncle Sam's procurement policies aren't the simplest things in the world, so the vendors who do the best are those who make it easy to buy from them and will bend over backwards before and after the sale to support their products.

Yep, you're right. When I get to this stage, I plan to hire a specialist to do this for me and who has done this before.

I can justify the cost, and here's why. I once had a web-based product that competed with Active Directory before Active Directory came out. It synchronized *nix, mainframe, and Windows accounts. It was built in ASP and VB6 along with Win32 API. (Yeah, ack, spit, spit -- this was years ago.) I had my own company then. It was Imetrus Corporation. I received several calls from the U.S. Military. Unfortunately, I wasn't GSA-listed and they had to turn me down. But then, Microsoft FINALLY released Active Directory and, even though my tool augmented Active Directory and enhanced it, it was still not popular enough to survive. In the end, I folded after nearly going bankrupt on building the business (my first startup) and MCI/Worldcom purchased my idea for pennies on the dollar. Another version of it ended up at a power company. I then squandered that money in other debts, very foolishly. So, if I had only gotten my act together and paid the cost for a GSA specialist, I would have perhaps been quite wealthy by now.

My next startup will be created with a lot more careful planning.

Offline
Joined: 2006-01-29
"supermike" wrote:

Yep, you're right. When I get to this stage, I plan to hire a specialist to do this for me and who has done this before.

I can justify the cost, and here's why. I once had a web-based product that competed with Active Directory before Active Directory came out. It synchronized *nix, mainframe, and Windows accounts. It was built in ASP and VB6 along with Win32 API. (Yeah, ack, spit, spit -- this was years ago.) I had my own company then. It was Imetrus Corporation. I received several calls from the U.S. Military. Unfortunately, I wasn't GSA-listed and they had to turn me down. But then, Microsoft FINALLY released Active Directory and, even though my tool augmented Active Directory and enhanced it, it was still not popular enough to survive. In the end, I folded after nearly going bankrupt on building the business (my first startup) and MCI/Worldcom purchased my idea for pennies on the dollar. Another version of it ended up at a power company. I then squandered that money in other debts, very foolishly. So, if I had only gotten my act together and paid the cost for a GSA specialist, I would have perhaps been quite wealthy by now.

My next startup will be created with a lot more careful planning.

Ouch! Squashed by the Redmond Giant. Well at least you have plenty of company in that camp!

Thats interesting you received calls from the US .mil. I'm actually an active duty member of the USAF (12 years now). But during the timeframe you're talking about I was still slaving away as a ground support equipment mechanic; I retrained in 2000, about the time Win2k and AD started. Luckily, I've been fortunate enugh to spend the majority of my career working with *nix systems.

It sounds like you've definitely got a workable plan going. Any idea on when release/launch day will be?

supermike's picture
Offline
Joined: 2006-02-17
"dotmil" wrote:

Any idea on when release/launch day will be?

The prototype has taken a long time to build in evenings and weekends. It's been about 10 years of learning and planning, and 4 of those were actually in coding. I'm now on the admin piece, which is then followed with the install piece. However, after a lot of thought, I'm still not satisfied with the internal code of it and want to rewrite some routines better. In some cases, I did things without regexps, but now I found that using regexps causes the code to run much faster. So I'm going to have to break down and do this. My ticketing system will not only work as a standalone project for those willing to use it as is, or they can customize it and consider it an SDK, much like SugarCRM and RequestTracker are that way today.

I'm probably looking at 6 months before I start forming an LLC, considering eCommerce options, etc.

I think it has widespread appeal. I can see all kinds of companies, government agencies, and non-profits using this sort of thing. And some may want it painless by purchasing a pre-installed hardware option.

Offline
Joined: 2006-01-29

Wow; you have put a lot of effort into this haven't you! Regexps can run faster, as long as the hardware is up to the task. I can't really offer much advice on other areas to possibly tweak without seeing your app.

BTW, any chance of having an open source version as well? It could help with future development of the app by getting more eyeballs involved.

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.